Ovarian carcinoma diagnosis: the clinical impact of 15 years of change.

ABSTRACT

BACKGROUND: Until recently ovarian carcinoma was considered to be a single disease, and treatment decisions were based solely on grade and pre- and postoperative tumour burden. New insights into molecular features, treatment response, and patient demographics led the scientific community to conclude that ovarian carcinoma histotypes are different disease entities. METHODS: In 2002, the pathology specimens from patients in a clinical trial were reviewed by an experienced gynaecopathologist (pathologist A) for translational research purposes. All cases were typed according to what were then current criteria. The identical cohort was now reassessed by the same expert pathologist and independently reviewed by another gynaecopathologist (pathologist B) applying WHO 2014 diagnostic criteria. Survival analyses were done based on the original as well as the new diagnoses, and historical biomarker study results were recalculated. RESULTS: Upon re-review, pathologist A rendered the same histotype diagnosis in only 54% of cases. In contrast, pathologists A and B independently rendered the same diagnosis in 98% of cases. Histotype was of prognostic significance when 2014 diagnoses were used, but was not prognostic using the original (2002) histotype diagnoses. CONCLUSIONS: Our study demonstrates a marked shift in ovarian carcinoma histotype diagnosis over the past 15 years. The new criteria are associated with a very high degree of interobserver reproducibility, allowing for treatment decisions based on histotype. Finally, biomarkers of putative prognostic significance were revealed to be primarily histotype-specific markers, confirming the critical importance of obtaining up-to-date diagnoses rather than accepting archival histotype data in clinical research.